Load the required modules for the project

library(tidyverse)
library(raster)          
library(sf)              
library(ggspatial)       
library(ggnewscale)     
library(ggsn)            
library(plotly)

Set the working directory

setwd(dirname(rstudioapi::getSourceEditorContext()$path))

Crime Rate and Unemployment Rate

1 Data

The data used was for this analysis is from three files. Each File contained data that needed to be joined in to a single coherent data frame that contained all the relevant information. The three files are the following:


    1. crime_and_incarceration_by_state.csv - holds information about crime statistics for each state and other data. It is composed of 18 different variables. Only a few will be used in the analysis. They are listed below. Variables, violent_crime_total and state_population, will be used to calculate the crime rate. The calculation can be found in Section: 3.0.5. “jurisdiction” and “year” column names will be changed in Section: 3.0.4. Additional information about the processing can be found in subsequent sections.

      The features that will be used from the file are:

      • Variable 1: violent_crime_total - the total amount of violent crime in the state
      • Variable 2: state_population - the total population of the state
      • Variable 3: jurisdiction - holds the state name.
      • Variable 4: year - holds the year the data was recorded


    1. unemployment_county.csv - This data will be processed in Section: 3.0.3. Additional information will be found in this section.

      The features that will be used from the file are:

      • Variable 1: Employed - the number of people employed in each of the counties
      • Variable 2: Unemployed - the number of people unemployed in each of the counties
      • Variable 3: Unemployment Rate - the unemployment rate for each of the counties


    1. tl_2019_us_state.shp - this is the shape file that will be used in the mapping the data in Sections: 3.2.1 and 3.2.2


Note: All other columns will not be used and will discarded during further processing.


2 Project Objectives

The objective of this project is to determine if there is a correlation between the rate of crime and unemployment. This project seeks to see if there is an observable correlation that be seen either graphically or seen in a mathematical supported way. There will be a discussion of the findings. The data topic that will covered in this project will be the temporal-spatial changes of unemployment rate in the contiguous USA.

3 Data Processing and Data Visualization

Data preprocessing:

Steps:

  • Read the data from the CSV files into individual data frames. Section: 3.0.1 .
  • Remove the parts of the United States that are not contiguous. Section: 3.0.2 .
  • Process the unemployment rate data. Section: 3.0.3
  • Process the crime rate data. Section: 3.0.4
  • Join relational tables. Section: 3.0.6
  • Save the final combined and cleaned data. Section: 3.0.7

3.0.1 Read in the data from the data files.

This code will read in data from the data files that were discussed in Section 1

`File Name` <- c('crime_and_incarceration_by_state.csv', 
                 'Murder Rates, States By Region_Full Data_data',
                 'tl_2019_us_state.shp')

file_df <- data.frame(`File Name`)

knitr::kable(file_df,col.names = c("File Name"), caption = "Files Used",)
Table 3.1: Files Used
File Name
crime_and_incarceration_by_state.csv
Murder Rates, States By Region_Full Data_data
tl_2019_us_state.shp 
# Read in the unemployment rate from the CSV file
Unemployrate <-  read_csv("data/unemployment_county.csv")

# Read in the Crime rate from the CSV file
Crimerate <- read_csv ("data/crime_and_incarceration_by_state.csv")

# Read the states shape file
States <- st_read("data/tl_2019_us_state/tl_2019_us_state.shp")
knitr::kable(head(Unemployrate, 10), caption = "Unemployment Rate")
Table 3.2: Unemployment Rate
County State Labor Force Employed Unemployed Unemployment Rate Year
Autauga County AL 24383 23577 806 3.3 2007
Baldwin County AL 82659 80099 2560 3.1 2007
Barbour County AL 10334 9684 650 6.3 2007
Bibb County AL 8791 8432 359 4.1 2007
Blount County AL 26629 25780 849 3.2 2007
Bullock County AL 3653 3308 345 9.4 2007
Butler County AL 9099 8539 560 6.2 2007
Calhoun County AL 54861 52709 2152 3.9 2007
Chambers County AL 15474 14469 1005 6.5 2007
Cherokee County AL 11984 11484 500 4.2 2007 
knitr::kable(head(Crimerate[ , 1:9], 10), 
             caption = "Crime and Incarceration by State Part 1")
Table 3.2: Crime and Incarceration by State Part 1
jurisdiction includes_jails year prisoner_count crime_reporting_change crimes_estimated state_population violent_crime_total murder_manslaughter
FEDERAL FALSE 2001 149852 NA NA NA NA NA
ALABAMA FALSE 2001 24741 FALSE FALSE 4468912 19582 379
ALASKA TRUE 2001 4570 FALSE FALSE 633630 3735 39
ARIZONA FALSE 2001 27710 FALSE FALSE 5306966 28675 400
ARKANSAS FALSE 2001 11489 FALSE FALSE 2694698 12190 148
CALIFORNIA FALSE 2001 157142 FALSE FALSE 34600463 212867 2206
COLORADO FALSE 2001 17278 FALSE FALSE 4430989 15492 158
CONNECTICUT TRUE 2001 17507 FALSE FALSE 3434602 11492 105
DELAWARE TRUE 2001 6841 FALSE FALSE 796599 4868 23
FLORIDA FALSE 2001 72404 FALSE FALSE 16373330 130713 874 
knitr::kable(head(Crimerate[ , 10: 17], 10),
             caption = "Crime and Incarceration by State Part 2")
Table 3.2: Crime and Incarceration by State Part 2
rape_legacy rape_revised robbery agg_assault property_crime_total burglary larceny vehicle_theft
NA NA NA NA NA NA NA NA
1369 NA 5584 12250 173253 40642 119992 12619
501 NA 514 2681 23160 3847 16695 2618
1518 NA 8868 17889 293874 54821 186850 52203
892 NA 2181 8969 99106 22196 69590 7320
9960 NA 64614 136087 1134189 232273 697739 204177
1930 NA 3555 9849 170887 28533 121360 20994
639 NA 4183 6565 95299 17159 65762 12378
420 NA 1156 3269 27399 5144 19476 2779
6641 NA 32867 90331 782517 176052 516548 89917 
knitr::kable(head(States, 10), caption = "States")
Table 3.2: States
REGION DIVISION STATEFP STATENS GEOID STUSPS NAME LSAD MTFCC FUNCSTAT ALAND AWATER INTPTLAT INTPTLON geometry
3 5 54 01779805 54 WV West Virginia 00 G4000 A 62266231560 489271086 +38.6472854 -080.6183274 MULTIPOLYGON (((-81.74725 3…
3 5 12 00294478 12 FL Florida 00 G4000 A 138947364717 31362872853 +28.4574302 -082.4091477 MULTIPOLYGON (((-86.38865 3…
2 3 17 01779784 17 IL Illinois 00 G4000 A 143779863817 6215723896 +40.1028754 -089.1526108 MULTIPOLYGON (((-91.18529 4…
2 4 27 00662849 27 MN Minnesota 00 G4000 A 206230065476 18942261495 +46.3159573 -094.1996043 MULTIPOLYGON (((-96.78438 4…
3 5 24 01714934 24 MD Maryland 00 G4000 A 25151726296 6979340970 +38.9466584 -076.6744939 MULTIPOLYGON (((-77.45881 3…
1 1 44 01219835 44 RI Rhode Island 00 G4000 A 2677787140 1323663210 +41.5974187 -071.5272723 MULTIPOLYGON (((-71.7897 41…
4 8 16 01779783 16 ID Idaho 00 G4000 A 214049897859 2391604238 +44.3484222 -114.5588538 MULTIPOLYGON (((-116.8997 4…
1 1 33 01779794 33 NH New Hampshire 00 G4000 A 23189198255 1026903434 +43.6726907 -071.5843145 MULTIPOLYGON (((-72.3299 43…
3 5 37 01027616 37 NC North Carolina 00 G4000 A 125925929633 13463401534 +35.5397100 -079.1308636 MULTIPOLYGON (((-82.41674 3…
1 1 50 01779802 50 VT Vermont 00 G4000 A 23874197924 1030383955 +44.0685773 -072.6691839 MULTIPOLYGON (((-73.31328 4…

3.0.2 Remove the parts of the United States that are not contiguous.


The states of Alaska, American Samoa, Northern Mariana Islands, Puerto Rico, US Virgin Islands, Hawaii, and Guam. The projects analysis will only focus on the contiguous United States or the mainland United States. Analysis will focus on the lower 48 states. 

Contiguous_state <- States %>% filter(STUSPS != "AK" & STUSPS != "AS" &
                                        STUSPS != "MP" & STUSPS != "PR" &
                                        STUSPS != "VI" & STUSPS != "HI" &
                                        STUSPS != "GU")

3.0.3 Process the unemployment rate data

The data will be filtered to remove Alaska and Hawaii from the data set. This analysis will only focus on the contiguous United States. It is not needed so it will be removed from the data. The data will be grouped by state and then by the Year in which the data was collected. Three variables will created. These variables are the following:

  • TotalForce: This variable will hold the total number of workers. This includes all workers both employed and unemployed.
  • Totalemployed: This variable will hold the total number of employed workers.
  • Totalunemployed: This variable will hold the total number of unemployed workers.
  • Meanrate: This variable will hold the mean rate of unemployment
Unemployrate <- Unemployrate %>% filter(State != 'AK' & State != "HI") %>%
  group_by(State, Year) %>% 
  summarise(Totalforce = sum(`Labor Force`), Totalemployed=sum(Employed),
            Totalunemployed=sum(Unemployed), Meanrate = mean(`Unemployment Rate`,
                                                             rm.na=TRUE))

The column in this data frame will need to have a column name changed from “State” to “STUSPS”. The years that will required will be also filtered from the data set. The years that are required for this project were from 2007 to 2014.

Unemployrate <- Unemployrate %>% rename("STUSPS" = "State") %>%
  filter(Year %in% c(2007:2014))

3.0.4 Process the Crime rate

In this step the crime rate will need to have two columns renamed using the rename() function. The two columns are jurisdiction and the year columns. The “jurisdiction” column will be changed to “STUSPS”. This will aid joining the frames in a later step. Changing “year” to “Year” will help keep the naming convention consistent among the data frames that are to be used in the final project.

Crimerate <- Crimerate %>% 
  rename("STUSPS" = "jurisdiction") %>%
  rename("Year" = "year") %>%
  filter(STUSPS != "FEDERAL" & STUSPS != "ALASKA" & STUSPS != "HAWAII") %>%
  filter(Year %in% c(2007:2014))

There will be a need to change the state names in the STUSPS column.

Crimerate$STUSPS <- state.abb[match(str_to_title(Crimerate$STUSPS), state.name)]

3.0.5 Calculate the crime rate

The crime rate was calculated using two columns from the Crimerate data frame. The columns were:

  • violent_crime_total: the total number of violent crime in the state
  • state_population: the population of the state
Crimerate <- Crimerate %>% 
  mutate(Crimerate=(violent_crime_total/state_population) * 100) %>%
  dplyr::mutate_if(is.numeric, round, 1)

3.0.6 Join relational tables

The data frames will be joined so all the data will be contained in one frame. Only unique columns will be included within the final data frame. From the joined data frames select columns that are relevant for final use in the creation of the final project.

CS_Erate <- right_join(Contiguous_state, Unemployrate, by= c("STUSPS"))

CS_Erate_Crate <- right_join(CS_Erate, Crimerate, by= c("STUSPS", "Year"))

CS_Erate_Crate1 <- CS_Erate_Crate %>% 
  select(REGION, STUSPS, NAME, Year, Meanrate,Crimerate) %>% 
  rename("Unemplyrate"="Meanrate")
knitr::kable(head(CS_Erate_Crate1, 10), caption = "Combined Data")
Table 3.3: Combined Data
REGION STUSPS NAME Year Unemplyrate Crimerate geometry
3 WV West Virginia 2007 5.138182 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2008 4.914546 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2009 8.801818 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2010 9.740000 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2011 8.985454 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2012 8.443636 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2013 7.716364 0.3 MULTIPOLYGON (((-81.74725 3…
3 WV West Virginia 2014 7.532727 0.3 MULTIPOLYGON (((-81.74725 3…
3 FL Florida 2007 4.186567 0.7 MULTIPOLYGON (((-86.38865 3…
3 FL Florida 2008 6.473134 0.7 MULTIPOLYGON (((-86.38865 3…

3.0.7 Save the final combined and cleaned data.

saveRDS(CS_Erate_Crate1, file = "CS_Erate_CrateCombined1.Rds")

3.1 EDA analysis

# Create a copy of the data frame
stats_df <- data.frame(CS_Erate_Crate1) %>% select(-geometry)

region_unemploy <- stats_df %>%
group_by(REGION) %>%
  summarise(
    `Region Mean` = mean(Unemplyrate),
    `Maximum Unemployment Rate` = max(Unemplyrate),
    `Minimum Unemployment Rate` = min(Unemplyrate),
    `Quantiles Unemployment` = list(round(quantile(Unemplyrate, type=1), 2)),
    `Standard Deviation` = sd(Unemplyrate),
  )

knitr::kable(region_unemploy, caption = "Regional Unemployment Statistics.", align = "cccc", digits = 2)
Table 3.4: Regional Unemployment Statistics.
REGION Region Mean Maximum Unemployment Rate Minimum Unemployment Rate Quantiles Unemployment Standard Deviation
1 7.02 10.54 3.50 3.50, 5.40, 7.19, 8.61, 10.54 1.83
2 6.52 14.12 2.87 2.87, 4.39, 5.97, 8.18, 14.12 2.53
3 8.04 13.27 3.43 3.43, 6.46, 7.75, 9.31, 13.27 2.33
4 7.69 13.81 2.92 2.92, 5.43, 7.58, 9.51, 13.81 2.71 
region_unemploy_box <- stats_df %>%
group_by(REGION) %>% ggplot(mapping=aes(x=REGION, y=Unemplyrate, fill=REGION))+
  geom_boxplot()+
  labs(colour="Year", y="Unemployment Rate", x="Region", 
       title="Unemployment Rate by Region") +
  theme(panel.background = element_blank(), text=element_text(size=16),
        plot.title=element_text(hjust=0.5, size=20))

ggplotly(region_unemploy_box)

Figure 3.1: Unemployment Rate by Region. 

region_year_unemploy <-stats_df %>%
group_by(Year, REGION) %>%
  summarise(
    `Region Mean` = mean(Unemplyrate),
    `Maximum Unemployment Rate` = max(Unemplyrate),
    `Minimum Unemployment Rate` = min(Unemplyrate),
    `Quantiles: 0%  25%  50%  75%  100%` = 
      list(round(quantile(Unemplyrate, type=1), 2)),
    `Standard Deviation` = sd(Unemplyrate),
  )

knitr::kable(region_year_unemploy, caption = "Regional Unemployment Statistics by Year and Region.",
             align = "cccc", digits = 2)
Table 3.5: Regional Unemployment Statistics by Year and Region.
Year REGION Region Mean Maximum Unemployment Rate Minimum Unemployment Rate Quantiles: 0% 25% 50% 75% 100% Standard Deviation
2007 1 4.52 5.34 3.50 3.50, 4.39, 4.46, 4.72, 5.34 0.49
2007 2 4.79 8.05 2.87 2.87, 3.73, 4.71, 5.33, 8.05 1.39
2007 3 5.03 7.15 3.43 3.43, 4.12, 5.00, 5.49, 7.15 1.13
2007 4 4.48 6.78 2.92 2.92, 3.43, 4.05, 5.69, 6.78 1.32
2008 1 5.58 7.20 3.84 3.84, 5.38, 5.54, 5.77, 7.20 0.89
2008 2 5.43 8.90 3.21 3.21, 3.60, 5.41, 6.26, 8.90 1.71
2008 3 6.10 8.38 3.69 3.69, 4.79, 6.21, 6.98, 8.38 1.38
2008 4 5.81 8.64 3.15 3.15, 4.60, 5.37, 7.22, 8.64 1.73
2009 1 8.27 10.22 6.17 6.17, 7.74, 8.22, 8.97, 10.22 1.21
2009 2 8.29 14.12 4.26 4.26, 5.17, 8.03, 9.95, 14.12 3.12
2009 3 9.80 13.27 6.48 6.48, 8.00, 8.80, 11.25, 13.27 2.18
2009 4 9.08 12.93 6.31 6.31, 6.49, 8.79, 11.67, 12.93 2.43
2010 1 8.55 10.54 5.82 5.82, 8.60, 8.78, 8.93, 10.54 1.45
2010 2 8.19 13.33 3.96 3.96, 5.25, 7.69, 10.16, 13.33 3.05
2010 3 10.20 13.15 7.16 7.16, 8.50, 9.74, 11.63, 13.15 1.82
2010 4 9.92 13.81 6.16 6.16, 8.48, 9.51, 12.09, 13.81 2.51
2011 1 8.13 10.40 5.38 5.38, 7.73, 8.42, 8.61, 10.40 1.59
2011 2 7.35 11.37 3.76 3.76, 5.21, 6.86, 9.09, 11.37 2.48
2011 3 9.60 12.58 6.20 6.20, 7.75, 9.31, 11.22, 12.58 1.81
2011 4 9.39 13.43 5.62 5.62, 7.53, 8.97, 11.59, 13.43 2.46
2012 1 7.85 9.79 5.40 5.40, 7.16, 8.18, 8.64, 9.79 1.58
2012 2 6.54 10.07 3.53 3.53, 4.85, 5.96, 8.07, 10.07 2.12
2012 3 8.62 11.12 5.52 5.52, 7.37, 8.53, 9.36, 11.12 1.55
2012 4 8.52 12.14 5.19 5.19, 6.29, 7.86, 10.23, 12.14 2.29
2013 1 7.19 8.70 4.90 4.90, 7.19, 7.68, 7.70, 8.70 1.40
2013 2 6.30 9.93 3.48 3.48, 4.41, 5.48, 7.75, 9.93 2.16
2013 3 8.03 10.13 5.65 5.65, 6.87, 7.72, 9.13, 10.13 1.29
2013 4 7.72 10.73 4.67 4.67, 5.53, 7.61, 9.06, 10.73 2.05
2014 1 6.03 7.22 4.19 4.19, 6.10, 6.19, 6.47, 7.22 1.07
2014 2 5.27 8.18 3.12 3.12, 3.99, 4.76, 6.21, 8.18 1.59
2014 3 6.95 8.91 4.81 4.81, 6.02, 7.07, 7.64, 8.91 1.11
2014 4 6.59 9.45 4.10 4.10, 4.74, 7.28, 7.81, 9.45 1.83 
region_year_unemploy_box <-
 ggplot(stats_df) + geom_boxplot(aes(x=REGION, y=Unemplyrate, fill=REGION)) + 
  facet_wrap(~Year, ncol=2) +
  labs(colour="Year", y="Unemployment Rate", x="Region", 
       title="Unemployment Rate by Year and Region") + 
  theme_classic() +
  theme(plot.title = element_text(hjust = 0.5, size=20), 
        text=element_text(size=16))

ggplotly(region_year_unemploy_box)

Figure 3.2: Unemployment Rate by Year and Region. 

region_crime <- stats_df %>%
group_by(REGION) %>%
  summarise(
    `Region Mean` = mean(Crimerate),
    `Maximum Crime Rate` = max(Crimerate),
    `Minimum Crime Rate` = min(Crimerate),
    `Quantiles: 0%  25%  50%  75%  100%` = list(quantile(Crimerate, type=1)),
    `Standard Deviation` = sd(Crimerate)
  )

knitr::kable(region_crime, caption = "Regional Crime Statistics", 
             align = "cccc", digits = 2)
Table 3.6: Regional Crime Statistics
REGION Region Mean Maximum Crime Rate Minimum Crime Rate Quantiles: 0% 25% 50% 75% 100% Standard Deviation
1 0.27 0.5 0.1 0.1, 0.2, 0.3, 0.4, 0.5 0.12
2 0.34 0.6 0.2 0.2, 0.3, 0.3, 0.4, 0.6 0.10
3 0.45 0.8 0.2 0.2, 0.3, 0.5, 0.5, 0.8 0.15
4 0.36 0.8 0.2 0.2, 0.2, 0.3, 0.4, 0.8 0.16 
region_crime_box <- stats_df %>%
group_by(REGION) %>% ggplot(mapping=aes(x=REGION, y=Crimerate, fill=REGION))+
  geom_boxplot() + 
   labs(colour="Year", y="Crime Rate", x="Region", 
       title="Crime Rate by Region") +
  theme(panel.background = element_blank(), 
        plot.title=element_text(hjust=0.5, size=20), text=element_text(size=16))

ggplotly(region_crime_box)

Figure 3.3: Crime Rate by Region. 

region_year_crime <-stats_df %>%
group_by(Year, REGION) %>%
  summarise(
    `Region Mean` = mean(Crimerate),
    `Maximum Crime Rate` = max(Crimerate),
    `Minimum Crime Rate` = min(Crimerate),
    `Quantiles: 0%  25%  50%  75%  100%` = list(quantile(Crimerate, type=1)),
    `Standard Deviation` = sd(Crimerate),
  )

knitr::kable(region_year_crime, caption = "Regional Crime Statistics by Year and Region.",
             align = "cccc", digits = 2)
Table 3.7: Regional Crime Statistics by Year and Region.
Year REGION Region Mean Maximum Crime Rate Minimum Crime Rate Quantiles: 0% 25% 50% 75% 100% Standard Deviation
2007 1 0.26 0.4 0.1 0.1, 0.1, 0.3, 0.4, 0.4 0.13
2007 2 0.37 0.6 0.2 0.2, 0.3, 0.3, 0.5, 0.6 0.13
2007 3 0.52 0.8 0.3 0.3, 0.3, 0.5, 0.7, 0.8 0.18
2007 4 0.43 0.8 0.2 0.2, 0.3, 0.4, 0.5, 0.8 0.18
2008 1 0.29 0.5 0.1 0.1, 0.2, 0.3, 0.4, 0.5 0.14
2008 2 0.36 0.5 0.2 0.2, 0.3, 0.3, 0.4, 0.5 0.10
2008 3 0.52 0.7 0.3 0.3, 0.3, 0.5, 0.7, 0.7 0.16
2008 4 0.39 0.7 0.2 0.2, 0.2, 0.3, 0.5, 0.7 0.19
2009 1 0.29 0.5 0.1 0.1, 0.2, 0.3, 0.4, 0.5 0.14
2009 2 0.34 0.5 0.2 0.2, 0.3, 0.3, 0.4, 0.5 0.11
2009 3 0.48 0.7 0.2 0.2, 0.3, 0.5, 0.6, 0.7 0.15
2009 4 0.36 0.7 0.2 0.2, 0.2, 0.3, 0.5, 0.7 0.17
2010 1 0.29 0.5 0.1 0.1, 0.2, 0.3, 0.4, 0.5 0.14
2010 2 0.32 0.5 0.2 0.2, 0.2, 0.3, 0.4, 0.5 0.11
2010 3 0.44 0.6 0.2 0.2, 0.3, 0.4, 0.5, 0.6 0.14
2010 4 0.35 0.7 0.2 0.2, 0.2, 0.3, 0.4, 0.7 0.16
2011 1 0.27 0.4 0.1 0.1, 0.2, 0.3, 0.4, 0.4 0.12
2011 2 0.31 0.4 0.2 0.2, 0.2, 0.3, 0.4, 0.4 0.08
2011 3 0.43 0.6 0.2 0.2, 0.3, 0.4, 0.5, 0.6 0.14
2011 4 0.34 0.6 0.2 0.2, 0.2, 0.3, 0.4, 0.6 0.15
2012 1 0.28 0.4 0.1 0.1, 0.2, 0.3, 0.4, 0.4 0.12
2012 2 0.33 0.5 0.2 0.2, 0.3, 0.3, 0.4, 0.5 0.10
2012 3 0.44 0.6 0.2 0.2, 0.3, 0.5, 0.5, 0.6 0.13
2012 4 0.34 0.6 0.2 0.2, 0.2, 0.3, 0.4, 0.6 0.15
2013 1 0.27 0.4 0.1 0.1, 0.2, 0.3, 0.3, 0.4 0.11
2013 2 0.33 0.5 0.2 0.2, 0.3, 0.3, 0.4, 0.5 0.08
2013 3 0.41 0.6 0.2 0.2, 0.3, 0.4, 0.5, 0.6 0.12
2013 4 0.34 0.6 0.2 0.2, 0.2, 0.3, 0.4, 0.6 0.15
2014 1 0.24 0.4 0.1 0.1, 0.2, 0.2, 0.3, 0.4 0.11
2014 2 0.32 0.4 0.2 0.2, 0.3, 0.3, 0.4, 0.4 0.06
2014 3 0.40 0.6 0.2 0.2, 0.3, 0.4, 0.5, 0.6 0.12
2014 4 0.34 0.6 0.2 0.2, 0.2, 0.3, 0.4, 0.6 0.15 
region_year_crime_box <-
 ggplot(stats_df) + geom_boxplot(aes(x=REGION, y=Crimerate, fill=REGION)) + 
  facet_wrap(~Year, ncol=2) +
  labs(colour="Year", y="Crimet Rate", x="Region", 
       title="Crime Rate by Year and Region") + 
  theme_classic() +
  theme(plot.title = element_text(hjust = 0.5, size=20), 
        text=element_text(size=16))

ggplotly(region_year_crime_box)

Figure 3.4: Crime Rate by Year and Region.

3.2 Data analytics method

The data visualizations that were produced for the project were the following:

  • A spatial map over the contiguous USA for the unemployment rate for the specific year 2014.
  • A spatial map over the contiguous USA for the crime rate for the specific year 2014.
  • Scatter plot for the data relationship between the unemployment rate and crime rate.
  • Time series plot for the four states for the unemployment rate
  • Time series plot for the four states for the crime rate

Data for the creation of the graphs is loaded from the RDS file that was created in a previous section of the project. The file is a “.Rds” the name of the file is:

  • CS_Erate_CrateCombined1.Rds

This file will read in using the readRDS(). The data found in this will then be used to create the plots that are found in this section of the project.

Read the cleaned data from the “.Rds” file.

all_info_from_RDS <- readRDS("CS_Erate_CrateCombined1.Rds")

3.2.1 A spatial map over the contiguous USA for the unemployment rate for the specific year 2014.

This is a map of the unemployment rate for the year 2014. This will be an interactive plot using the plot_ly function to create it.

The only year that will plotted on this time series plot will be for the year 2014. This data will be filtered from the all_info_from_RDS.

Note: This step could have been done using a pipe, but this makes it easier to see what is going on.

info_for_year_2014 <- all_info_from_RDS %>% filter(all_info_from_RDS$Year == 2014)

Using the info_for_year_2014 data frame a graph of the contiguous United States will be created showing unemployment rate as a layer on the graph.

sp1 <- ggplot(data=info_for_year_2014) + 
  geom_sf(data= info_for_year_2014$geometry, 
          aes(fill=info_for_year_2014$Unemplyrate, 
              text=paste("State: ",info_for_year_2014$NAME ,
                         "\nUnemployment Rate: ", 
                         round(info_for_year_2014$Unemplyrate, 2 )))) + 
  xlab("Longitude") +
  ylab("Latitude") +
  guides(fill=guide_legend(title= "Unemployment Rate for 2014")) + 
  labs(title = "Unemployment Rate Over Contiguous USA ",
       subtitle = "Unemployment Color Coded by State",
       caption = "Data source: Unknown") +
  scalebar(data= info_for_year_2014, location="bottomleft", dist= 500, st.size=2,
           dist_unit = "km", transform= TRUE, model= "WGS84", st.dist=0.04) +
  annotation_north_arrow(location = "br", which_north = "true", 
                         style = north_arrow_fancy_orienteering) +
  theme(panel.background = element_blank(), legend.position = "right", 
        plot.title = element_text(hjust = 0.5, size=20),
        plot.subtitle = element_text(hjust = 0.5, size=16),
        text=element_text(size=16))

sp1
A spatial map over the contiguous USA for the unemployment rate for the year 2014

Figure 3.5: A spatial map over the contiguous USA for the unemployment rate for the year 2014

3.2.2 A spatial map over the contiguous USA for the crime rate for the specific year 2014.

Using the info_for_year_2014 data frame a graph of the contiguous United States will be created showing crime rate as a layer on the graph.

ggplot(data=info_for_year_2014) + 
  geom_sf(data= info_for_year_2014$geometry, 
          aes(fill=info_for_year_2014$Crimerate)) + 
  xlab("Longitude") +
  ylab("Latitude") +
  guides(fill=guide_legend(title= "Crime Rate for 2014")) + 
  labs(title = "Crime Rate Over Contiguous USA ",
       subtitle = "Crime Rate Color Coded by State",
       caption = "Data source: Unknown") + 
  scalebar(data= info_for_year_2014, location="bottomleft", dist= 500, st.size=2,
           dist_unit = "km", transform= TRUE, model= "WGS84", st.dist=0.04) +
  annotation_north_arrow(location = "br", which_north = "true", 
                         style = north_arrow_fancy_orienteering) +
  theme(panel.background = element_blank(), 
        plot.title = element_text(hjust = 0.5, size=20),
        plot.subtitle = element_text(hjust = 0.5, size=16),
        text=element_text(size=16))
Spatial map over the contiguous USA for the crime rate for the year 2014

Figure 3.6: Spatial map over the contiguous USA for the crime rate for the year 2014

Scatter plot for the data relationship between the unemployment rate and crime rate.

Creates a scatter plot using crime rate (x-axis) and unemployment rate (y-axis).

fig <- plot_ly(data= info_for_year_2014, x= ~Crimerate, y= ~Unemplyrate,
                color= ~REGION) %>%
  add_markers() %>%
  layout(title="<b>Unemployment Rate and Crime Rate for 2014 </b>", 
         margin=list(b = 10, l= 10)) %>%
  layout(xaxis=list(title= "<b>Crime Rate Per 100,000 People</b>"),
         yaxis=list(title="<b>Unemployment Rate Per 100 People </b>"),
         legend=list(title=list(text='<b> Region </b>'),
                     showlegend=TRUE)) %>%
  layout(xaxis=list(titlefont= list(size= 14)),
         yaxis=list(titlefont= list(size= 14)))
   

fig

Figure 3.7: Scatter plot for the data relationship between the unemployment rate and crime rate

4) Time series plot for the four states for the unemployment rate.

This will be an interactive plot of the unemployment rate for four states:

  • California
  • Idaho
  • Illinois
  • Indiana

Steps to create the time series plot:

  • Data will be filtered from the all_info_from_RDS data frame and a new data frame will be created. Section: 3.3
  • The new data frame created is four_states_year_2014. Section: 3.3
  • Create the unemployment rate time series plot. Section: 3.4
  • Create the crime rate time series plot.Section: 3.5

Section 3.3 data filtered from the all_info_from_RDS data frame and a new data frame will be created. A vector of states was created to form the list of states that were to plotted on the graph. These states will be used for this time series plot and the one that follows.

3.3 Data will be filtered from the all_info_from_RDS data frame and a new data frame will be created.

states <- c("California", "Idaho", "Illinois", "Indiana") 
four_states_year_2014 <- all_info_from_RDS %>% filter(NAME %in% states)

stats_df <-  as.data.frame(four_states_year_2014)

3.4 Create the unemployment rate time series plot.

une <- plot_ly(data=stats_df, x= ~as.factor(Year), y= ~Unemplyrate,color= ~NAME) %>%
  filter(NAME %in% states) %>%
  group_by(NAME) %>%
  add_lines() %>%
  layout(title="<b>Unemployment Rate Changes by Year</b>",  
         xaxis=list(title= "<b>Year</b>"),
         yaxis=list(title="<b>Unemployment Rate</b>"), 
         legend=list(title=list(text='<b> State </b>'), showlegend=TRUE))

une

Figure 3.8: Unemployment rate time series plot.

3.5 Create the crime rate time series plot.

Note: To better see the crime rate for California select it from the legend on the right of the plot.

cr <- plot_ly(data=stats_df, x= ~as.factor(Year), y= ~Crimerate, color= ~NAME) %>%
  filter(NAME %in% states) %>%
  group_by(NAME) %>%
  add_lines() %>%
  layout(title="<b>Crime Rate Changes by Year</b>",  
         xaxis=list(title= "<b>Year</b>"),
         yaxis=list(title="<b>Crime Rate</b>"), yaxis=list(range(c(0, .7))),
         legend=list(title=list(text='<b> State </b>'), showlegend=TRUE))

  
cr

Figure 3.9: Crime rate time series plot.

4 Discussion and Conclusion

This project proposed to investigate if there was a relationship between changes in the unemployment rate and the crime rate. This project resulted in numeric and graphical data that can be used to answer the following questions:

  1. Does the employment rate change over time and in a specific region of the country?
  2. Can we discern a pattern or a correlation between the changes in the unemployment rate and the geography in which the data has been taken?

While it might be self-evident that an increase in unemployment would increase the crime rate, we can use the analysis that resulted from the graphs and the numerical data to see if there is a correlation. Many methods were employed to see if there was a correlation between the two.

Examining the statistics found in Section 3.1 EDA Analysis, we can see in tables that show some statistical measurements for the data. The tables and graphs show the quantiles, the mean, maximum, minimum, and any outliers for the available data. Table 3.4 Regional Unemployment Statistics. Shows the grouped data for all the years within the dataset. In Figure 3.1 Unemployment Rate by Region. Based on the boxplots regions 3 and 4 seem to have an extreme range in the unemployment rate for the period.

If we examine the data for each year as given in Table 3.5 Regional Unemployment Statistics by Year and Region, regions 3 and 4 usually have the highest mean unemployment rate over the period. To better illustrate the measurements that were given in the table, it may be helpful to look at a boxplot of the data over time. Using a faceted boxplot we can visualize the data for the regions for each year that there is data. Studying Figure 3.2 Unemployment Rate by Year and Region we can see that regions 2 and 4 do show the extremes in the unemployment rate.

The next item to look at is the crime rate in aggregate. Looking at the data over time we can see that the regions with the highest mean change in crime rate are regions 3 and 4. In Table 3.6 Regional Crime Statistics we can see the statistics for aggregated crime rate. These statistics seem to correspond with the high unemployment rates in regions 3 and 4, as discussed previously. To visualize this data we can refer to Figure 3.3 Crime Rate by Region. In this boxplot, we can see the range in the rate of change in the crime rate for regions. Regions 3 and 4 have extreme changes in crime for the data.

Referring to Table 3.7 Regional Crime Statistics by Year and Region shows the crime rate statistics by year and the regions. Over the period regions 3 and 4 continue to show the highest change in crime. To make it easier to understand it may be beneficial to look at the boxplots of the crime rate change over the year 2007 to 2014.

This can be seen in Figure 3.4 Crime Rate by Year and Region the regions show consistently high changes in the crime rate. If we look at a spatial map for a year, is there a suggestion of any correlation between the crime rate and the unemployment rate? The first spatial map to be examined is for the unemployment this map can be found in Figure 3.5. The highest rate of unemployment for 2014 is about 9% for the states of Mississippi and Arizona. The state of Mississippi is in Region 3 and Arizona is in Region 4.

What does this mean for the crime rate for the states looking at Figure 3.6? The rates for these states in 2014 were in the range of 0.1 for Mississippi and 0.4 for the state of Arizona. While there does seem to be some discrepancy in terms of the crime rate change, this may be due to the time frame that was chosen to examine for the spatial map.

Sometimes it may be beneficial to look at the data points as they relate to each other. This can be accomplished using the scatterplot. In Figure 3.7, there is a scatterplot of the data for the year 2014, based on this plot there does not seem to be a linear relation between the unemployment rate and the crime rate. There does not seem to be a linear correlation between the data points as indicated by the graph (Question Video: Identifying the Linear Correlation from the Scattergraph, Nagwa, n.d.).

When looking at a few examples from the selection of states we can further look for a trend in in crime and the unemployment rate. Using the states of California, Idaho, Illinois, and Indiana we can see the unemployment rates from 2007 to 2014 (Please refer to Figure 3.8 Unemployment Rate Changes by Year and Figure Crime Rate Changes by Year 3.9. (Please note it may be beneficial to toggle the different states on and off by selecting them from the legend located on the right side of the graph.)

We can see that California had the highest unemployment in the years 2010 through 2014. California had a change in the crime rate that was greater than the other states that were listed in the graph. One thing to note is Illinois had a spike in the unemployment rate in 2013 and did not see a commensurate change in the crime rate as indicated by no change in the crime rate shown in Figure 3.9.

To look at this relationship in a numerical manner we can look at the correlation coefficient. If we look at the correlation coefficient 0.17. The value of 0.17 indicates that there is almost no correlation between the variables to indicate that there is a meaningful correlation between crime and unemployment rates. This value is very close to the value zero. For there to be a correlation between the variables the value needs to fall closer to either -1 or to 1. Values closer to -1 indicate a negative correlation indicating that the variables change in a negative relation to each other. If one of the variables increases the other will decrease and vice versa (Soetewey, 2020).

While looking at data utilizing different methods. It seems that we can determine that there may be a possible correlation based on the graphs, but using a scatterplot there does not seem to be a correlation between the unemployment rate and the crime rate. There is a natural tendency to think that as unemployment increases the crime rate will also increase and it seems to be true.

Upon further research into the subject of crime rate and unemployment, there is a correlation between the two. This correlation depends on the type of crime that is being examined. When there is high unemployment the type of crime will vary. Crimes that involve possible quality of living may be only marginally affected. For instance, larceny, burglary, and robbery have a positive relationship. Crimes that may not affect the quality of living like car theft do not show a positive correlation but illustrate a negative relationship. This is not to say that car theft does not go up if given the right circumstances (Fallahi & Rodríguez, 2014). For instance, car theft will go up if there is a period of expansion within the economy. This is logical since there will be a market for the cars and the parts that they contain.

For further investigative purposes, it might be beneficial to look into the regions that were covered in the data to see what the prevalent form of employment is and see if there are any similarities in the type of work that may cause such high rates of unemployment.

5 References

Fallahi, F., & Rodríguez, G. (2014). Link between unemployment and crime in the US: A Markov-Switching approach. ScienceDirect.

Question Video: Identifying the Linear Correlation from the Scattergraph, Nagwa. (n.d.). Identifying the Linear Correlation from the Scattergraph [Video]. Nagwa. https://www.nagwa.com/en/videos/909167139353/

Soetewey, A. (2020, May 28). Correlation coefficient and correlation test in R. Stats and R. Retrieved February 13, 2025, from https://statsandr.com/blog/correlation-coefficient-and-correlation-test-in-r/